Not sure if you can see them in this video, this was a mast experiment which used one motor and two cables.
http://www.youtube.com/watch?v=zGf9OHUmHHw
At the ends of each cable there was a spring, those springs do take up a bit of slack but primarily they handle shock loads at motor start or direction changes. The arm-spinner wheel is a sandwich made from 5 Home Depot bucket lids, you can't see the spring on that cable, it lives in a curved slot on the back side of the wheel, but you can see the spring in the return cable, after the arm comes up its just above the spinner wheel.
The cables are wound on two spools which spin on a common shaft, but the windings go in different directions. If the two spools were locked to the shaft, then turning the shaft would wind up one spool and unwind the other spool the same number of turns. But the two spools can not be fixed permanently together because the diameters of the wound cable are changing, one is winding when one is unwinding. So in between the two inner sheaves of the spools, I put a "clutch disc", just a circle cut out of a sheet of Scotchbrite pad. The two spools would try to move away from each other, I tensioned them up against each other with a couple of drawer slide brackets. The slide brackets mounted to the base which holds the spools, the little wheels on the slide brackets push on the inner sheaves, there is a screw adjuster which squeezes the Scotchbrite to an appropriate amount of friction.
Two spool drive designs evolved. The first one allowed the shaft to slide back and forth, the shaft had two pins, when it was all the way in one direction the pin on that end would engage the nub on the outer sheave of that spool, as described above this would drive that spool and the other spool gets dragged along by the Scotchbrite. The ability to play with one spool was handy during assembly and debug, you could move the shaft to a point where neither pin engaged, lock one spool with Vise-grips, and then a human could pull a cable, or wind up a cable, as needed without bothering the other spool. Hopefully this photo is not too big...
After it was all working, the second design just moved the spool drive pins inward, allowing almost a full rotation of a spool before the pin bumped into the nub. So that means there was almost one turn of differential motion available during the up/down travel. If you use larger diameter spools, the windup differential is less of an issue.